微型电脑应用

为了挖掘网络节点数据，实现网络关键节点挖掘，提出多层时序有偏PageRank算法的网络中关键节点数据挖掘方法。运用时间层之间与层内节点之间的连接关系描述时序网络，以此为基础，采用节点层间相似性的超邻接矩阵(SSAM)方法构建多层时序网络模型。在SSAM多层时序网络模型中，基于有偏随机游走过程计算网络节点的转移概率矩阵，确定游走者下一个跳转的邻近节点，采用PageRank方法计算转移概率矩阵所确定跳转节点的KeyRank值，依据KeyRank值完成多层时序网络中跳转节点的重要度排序，实现多层时序网络中关键节点挖掘。实验结果表明，所提方法能够考虑时间层之间的相似性与差异性，提高关键节点挖掘的准确性。

This paper proposes the data mining method of key nodes in network based on the multi-layer temporal biased PageRank algorithm to mine network node data, so as to realize network key nodes mining. The temporal network is described by the connection relationship between the time layers and the nodes in the layer, based on which a multi-layer temporal network model is built by similarity-based supra-adjacency matrix(SSAM) method. In the SSAM multi-layer temporal network model, the transition probability matrix of the network node is calculated based on the biased random walk process, and the next jump neighbor node of the walker is determined. The PageRank method is used to calculate the KeyRank value of the jump node determined by the transition probability matrix, and the importance ranking of the jump node in the multi-layer temporal network is completed according to the KeyRank value, so as to realize the key node mining in the multi-layer temporal network. The experiment results show that the proposed method can consider the similarity and difference between time layers and improve the accuracy of key node mining.